Fig. 4

Immunity and DNA damage response were the prognostic factors of radiotherapy. A High radiocurability patients showed favorable OS in HNSC. B Word clouds of GSEA between high-low radiocurability patients in HNSC. C & D GSEA between high-low radiocurability patients in HNSC. E–H Survival curve, word clouds, and GSEA of LUAD. I-K Pan-cancer analysis of GSEA revealed that immunity, DDR, and angiogenesis influenced radiocurability. The value is the normalized enrichment score of GSEA for differentially expressed genes between high-low radiocurability patients. L The prognostic value of HRD scores and mutation counts via traversal method (see traversal method in Additional file 1). M The cut-off-HR scatters of mutation counts in CESC. N Survival curve of high-low mutation counts in CESC. O The prognostic value of immune infiltration via traversal method. Hierarchical clustering of SCM Q can effectively characterize the similarity of immune infiltration matrix P. R & S ANN-SCGP model performed well in both training (TCGA LUAD) and testing (GSE68465) sets using immune infiltration matrix as inputs. RT, radiotherapy; SCM, selectively connected matrix; Mut, mutation; HR, hazard ratio